Re: [tied] Circumflex accent

On Sun, 20 Apr 2003 14:16:13 +0000, Glen Gordon
<glengordon01@...> wrote:

>
>Miguel:
>>An opposition between circumflex and acute is attested independently in
>>Greek, Balto-Slavic and Germanic
>
>Alright. Then this hiatus thing occured immediately after
>Anatolian was a distinct and seperate dialect _during_ the
>breakup of Indo-European.

Why after the break-off of Anatolian? Anatolian has innovated and
doesn't offer much evidence for the ancient shape of the o-stems, but
what we have supports a distinction between Dsg. -o:(i) (> -a, the
Hittite allative) and Lsg. -oi (> -i).

>Thus postIE *-oei as you had said in cases where the *o is
>accented, otherwise postIE *-o:i as in ulkWo:i. However, in
>_common_ IE, I'd be more inclined to write dative *-oi (accented
>in *yug-oi but unaccented in *ulkWo-i) and a simple locative in
>*-i.

Why *-oi? The dative is clearly *-éi, the locative *-i (< unstressed
*-ei).

>At any rate, in both our scenarios, hiatus must be considered
>very recent.

I actually think it's quite ancient. It *must* predate zero grade,
otherwise we would have had o-stem Gsg. +-esyo instead of *-osyo; the
thematic vowel was lengthened (/e/ > /o/) before the unstressed vowel
/e/, before it was reduced by zero grade. We must have had a stage
-%-esyó or -%'-esyo (and likewise D. -%-é, L. -%'-e, I. -%-ét, Ab.
-%'-ot), with hiatus, _before_ the peculiar Ablaut of the thematic
vowel and _before_ zero grade. That's pretty ancient in my book.


=======================
Miguel Carrasquer Vidal
mcv@...